The prevalence of diabetes among inpatients in medical wards, surgery and intensive care units in Italy is approximatively 12-25%. The management of in hospital diabetes and hyperglycemia is complex and requires a specific training for physicians and nurses in non-specialist settings. The overall project aims at the implementation of a "best practice" model of care for hospitalized diabetic patients in non-specialist settings.

The project aims at defining a model of care for hospitalized diabetic patients in non-specialist settings based on the national and international recommendations. This is a cluster randomized study to evaluate the efficacy of a structured education training program for physicians and nurses in improving an ad hoc composite performance score of diabetes management and some clinical outcomes. The study will be carried out in 6 hospitals of the Emilia-Romagna region in Italy and it will involve 3 different wards (internal medicine, intensive care unit, surgery). For each hospital one intervention and two controls wards will be randomly identified.

At baseline the clinical management of patients with stress hyperglycemia/diabetes will be assessed in all participating wards over a 3-month period. Then, a structured educational program regarding the management of stress hyperglycemia/diabetes in the inpatient setting will be carried out only in the intervention wards. This program will last 2 months. Health care professionals also will be trained in the regular use of appropriate management tools (hyperglycemia dedicated standard operation procedures, audits, etc.).

Three months after the end of the educational program, the 3-month assessment of clinical management of patients with stress hyperglycemia/diabetes will be repeated in all participating wards.

The assessment of clinical management of stress hyperglycemia/diabetes will be based on a set of clinical performance indicators concerning 5 major domains (admission assessment, glucose monitoring, medical therapy, management of consults, management of discharge). These indicators will be used to define a composite performance score of appropriateness and efficacy. The highest score (1) will be assigned when the indicator is monitored, the lowest (0) when the indicator will be not detected, according to the scoring method previously reported by Rossi et al. Therefore, the score of each domain will range between 0 and the number of indicators used to assess appropriateness and efficacy of that domain.

Domain 1 - Initial assessment: score 0-5

• records of admission glycemia
• records of fasting plasma glucose
• records of HbA1c
• appropriate diagnosis of stress hyperglycemia/diabetes mellitus according to ADA
• records of history of presence/absence of pharmacological diabetes therapy

Domain 2 - Glucose monitoring: score 0-4

• appropriate glucose monitoring: at least 80% of three daily preprandial glycemia or, in critically ill patients, monitoring according to the specific algorithm used in the ward
• presence and use of specific forms for glucose records
• presence of ward standing orders to recognize and correct hypoglycemia
• presence of ward standing orders to monitor and manage glucose in critically ill patients (Critically ill patients are defined as patients who require intensive care for acute critical illnesses, such as myocardial infarction, stroke, septic shock or severe respiratory failure, requiring intensive or semi-intensive therapy and, as a rule, not taking food orally during the first 24-72 hours)

Domain 3 - Medical therapy: score 0-2

• records of time and dose of insulin therapy
• interruption of Metformin therapy when indicated (correct indications to discontinuation of Metformin therapy were: any critical illness, acute renal, cardiac and/or respiratory failure, surgery or iodinated contrast agents administration.)

Domain 4 - Consults: score 0-2

• request and records of diabetes specialist consult
• request and records of nursing consult

Domain 5 - Management of discharge: score 0-3

• planning of diabetes follow-up visit after hospitalization
• patient education for home blood glucose self-monitoring
• patient education for all other aspects of diabetes management

The sum of the scores of each domain from 1 to 5 is the composite performance score.

The following clinical outcomes will be recorded both at baseline and 3 months after the end of the educational program in all participating wards.:

• all hypoglycemic events (blood glucose ≤ 70 mg/dl)
• severe hypoglycemic events (blood glucose ≤ 40 mg/dl with or without clouding of consciousness)
• difference between plasma glycemia at admission and mean plasma glucose during the last-48-hours before discharge
• achievement of glycemic goals: 4 consecutive blood glucose levels ≤ 130 mg/dl (preprandial) or ≤ 180 mg/dl (postprandial), or 4 consecutive blood glucose levels between 140 and 180 mg/dl in critically ill patients
• survival
• discharge condition: a score of 0 will be assigned in case of death or transfer to a higher intensity ward and score 1 in all other cases, i.e. home discharge or transferral to a lower intensity ward.

Inclusion Criteria:

• patients with type 2 or type 1 diabetes or stress hyperglycemia

Exclusion Criteria:

• age <18,
• patients hospitalized for diabetic complications: ketoacidosis, hyperosmolar syndrome, hypoglycemic coma, pregnant diabetics or gestational diabetes